1CREATE FUNCTION() SQL Commands CREATE FUNCTION()
2
3
4
6 CREATE FUNCTION - define a new function
7
8
10 CREATE [ OR REPLACE ] FUNCTION
11 name ( [ [ argmode ] [ argname ] argtype [, ...] ] )
12 [ RETURNS rettype ]
13 { LANGUAGE langname
14 | IMMUTABLE | STABLE | VOLATILE
15 | CALLED ON NULL INPUT | RETURNS NULL ON NULL INPUT | STRICT
16 | [ EXTERNAL ] SECURITY INVOKER | [ EXTERNAL ] SECURITY DEFINER
17 | AS 'definition'
18 | AS 'obj_file', 'link_symbol'
19 } ...
20 [ WITH ( attribute [, ...] ) ]
21
22
24 CREATE FUNCTION defines a new function. CREATE OR REPLACE FUNCTION
25 will either create a new function, or replace an existing definition.
26
27 If a schema name is included, then the function is created in the spec‐
28 ified schema. Otherwise it is created in the current schema. The name
29 of the new function must not match any existing function with the same
30 argument types in the same schema. However, functions of different
31 argument types may share a name (this is called overloading).
32
33 To update the definition of an existing function, use CREATE OR REPLACE
34 FUNCTION. It is not possible to change the name or argument types of a
35 function this way (if you tried, you would actually be creating a new,
36 distinct function). Also, CREATE OR REPLACE FUNCTION will not let you
37 change the return type of an existing function. To do that, you must
38 drop and recreate the function. (When using OUT parameters, that means
39 you can't change the names or types of any OUT parameters except by
40 dropping the function.)
41
42 If you drop and then recreate a function, the new function is not the
43 same entity as the old; you will have to drop existing rules, views,
44 triggers, etc. that refer to the old function. Use CREATE OR REPLACE
45 FUNCTION to change a function definition without breaking objects that
46 refer to the function.
47
48 The user that creates the function becomes the owner of the function.
49
51 name The name (optionally schema-qualified) of the function to cre‐
52 ate.
53
54 argmode
55 The mode of an argument: either IN, OUT, or INOUT. If omitted,
56 the default is IN.
57
58 argname
59 The name of an argument. Some languages (currently only
60 PL/pgSQL) let you use the name in the function body. For other
61 languages the name of an input argument is just extra documenta‐
62 tion. But the name of an output argument is significant, since
63 it defines the column name in the result row type. (If you omit
64 the name for an output argument, the system will choose a
65 default column name.)
66
67 argtype
68 The data type(s) of the function's arguments (optionally schema-
69 qualified), if any. The argument types may be base, composite,
70 or domain types, or may reference the type of a table column.
71
72 Depending on the implementation language it may also be allowed
73 to specify ``pseudotypes'' such as cstring. Pseudotypes indi‐
74 cate that the actual argument type is either incompletely speci‐
75 fied, or outside the set of ordinary SQL data types.
76
77 The type of a column is referenced by writing tablename.column‐
78 name%TYPE. Using this feature can sometimes help make a func‐
79 tion independent of changes to the definition of a table.
80
81 rettype
82 The return data type (optionally schema-qualified). The return
83 type may be a base, composite, or domain type, or may reference
84 the type of a table column. Depending on the implementation
85 language it may also be allowed to specify ``pseudotypes'' such
86 as cstring. If the function is not supposed to return a value,
87 specify void as the return type.
88
89 When there are OUT or INOUT parameters, the RETURNS clause may
90 be omitted. If present, it must agree with the result type
91 implied by the output parameters: RECORD if there are multiple
92 output parameters, or the same type as the single output parame‐
93 ter.
94
95 The SETOF modifier indicates that the function will return a set
96 of items, rather than a single item.
97
98 The type of a column is referenced by writing tablename.column‐
99 name%TYPE.
100
101 langname
102 The name of the language that the function is implemented in.
103 May be SQL, C, internal, or the name of a user-defined procedur‐
104 al language. For backward compatibility, the name may be
105 enclosed by single quotes.
106
107 IMMUTABLE
108
109 STABLE
110
111 VOLATILE
112 These attributes inform the query optimizer about the behavior
113 of the function. At most one choice may be specified. If none of
114 these appear, VOLATILE is the default assumption.
115
116 IMMUTABLE indicates that the function cannot modify the database
117 and always returns the same result when given the same argument
118 values; that is, it does not do database lookups or otherwise
119 use information not directly present in its argument list. If
120 this option is given, any call of the function with all-constant
121 arguments can be immediately replaced with the function value.
122
123 STABLE indicates that the function cannot modify the database,
124 and that within a single table scan it will consistently return
125 the same result for the same argument values, but that its
126 result could change across SQL statements. This is the appropri‐
127 ate selection for functions whose results depend on database
128 lookups, parameter variables (such as the current time zone),
129 etc. Also note that the current_timestamp family of functions
130 qualify as stable, since their values do not change within a
131 transaction.
132
133 VOLATILE indicates that the function value can change even
134 within a single table scan, so no optimizations can be made.
135 Relatively few database functions are volatile in this sense;
136 some examples are random(), currval(), timeofday(). But note
137 that any function that has side-effects must be classified
138 volatile, even if its result is quite predictable, to prevent
139 calls from being optimized away; an example is setval().
140
141 For additional details see in the documentation.
142
143 CALLED ON NULL INPUT
144
145 RETURNS NULL ON NULL INPUT
146
147 STRICT CALLED ON NULL INPUT (the default) indicates that the function
148 will be called normally when some of its arguments are null. It
149 is then the function author's responsibility to check for null
150 values if necessary and respond appropriately.
151
152 RETURNS NULL ON NULL INPUT or STRICT indicates that the function
153 always returns null whenever any of its arguments are null. If
154 this parameter is specified, the function is not executed when
155 there are null arguments; instead a null result is assumed auto‐
156 matically.
157
158 [EXTERNAL] SECURITY INVOKER
159
160 [EXTERNAL] SECURITY DEFINER
161 SECURITY INVOKER indicates that the function is to be executed
162 with the privileges of the user that calls it. That is the
163 default. SECURITY DEFINER specifies that the function is to be
164 executed with the privileges of the user that created it.
165
166 The key word EXTERNAL is allowed for SQL conformance, but it is
167 optional since, unlike in SQL, this feature applies to all func‐
168 tions not only external ones.
169
170 definition
171 A string constant defining the function; the meaning depends on
172 the language. It may be an internal function name, the path to
173 an object file, an SQL command, or text in a procedural lan‐
174 guage.
175
176 obj_file, link_symbol
177 This form of the AS clause is used for dynamically loadable C
178 language functions when the function name in the C language
179 source code is not the same as the name of the SQL function. The
180 string obj_file is the name of the file containing the dynami‐
181 cally loadable object, and link_symbol is the function's link
182 symbol, that is, the name of the function in the C language
183 source code. If the link symbol is omitted, it is assumed to be
184 the same as the name of the SQL function being defined.
185
186 attribute
187 The historical way to specify optional pieces of information
188 about the function. The following attributes may appear here:
189
190 isStrict
191 Equivalent to STRICT or RETURNS NULL ON NULL INPUT.
192
193 isCachable
194 isCachable is an obsolete equivalent of IMMUTABLE; it's
195 still accepted for backwards-compatibility reasons.
196
197 Attribute names are not case-sensitive.
198
200 Refer to in the documentation for further information on writing func‐
201 tions.
202
203 The full SQL type syntax is allowed for input arguments and return
204 value. However, some details of the type specification (e.g., the pre‐
205 cision field for type numeric) are the responsibility of the underlying
206 function implementation and are silently swallowed (i.e., not recog‐
207 nized or enforced) by the CREATE FUNCTION command.
208
209 PostgreSQL allows function overloading; that is, the same name can be
210 used for several different functions so long as they have distinct
211 argument types. However, the C names of all functions must be differ‐
212 ent, so you must give overloaded C functions different C names (for
213 example, use the argument types as part of the C names).
214
215 Two functions are considered the same if they have the same names and
216 input argument types, ignoring any OUT parameters. Thus for example
217 these declarations conflict:
218
219 CREATE FUNCTION foo(int) ...
220 CREATE FUNCTION foo(int, out text) ...
221
222
223 When repeated CREATE FUNCTION calls refer to the same object file, the
224 file is only loaded once. To unload and reload the file (perhaps during
225 development), use the LOAD [load(7)] command.
226
227 Use DROP FUNCTION [drop_function(7)] to remove user-defined functions.
228
229 It is often helpful to use dollar quoting (see in the documentation) to
230 write the function definition string, rather than the normal single
231 quote syntax. Without dollar quoting, any single quotes or backslashes
232 in the function definition must be escaped by doubling them.
233
234 To be able to define a function, the user must have the USAGE privilege
235 on the language.
236
238 Here are some trivial examples to help you get started. For more infor‐
239 mation and examples, see in the documentation.
240
241 CREATE FUNCTION add(integer, integer) RETURNS integer
242 AS 'select $1 + $2;'
243 LANGUAGE SQL
244 IMMUTABLE
245 RETURNS NULL ON NULL INPUT;
246
247
248 Increment an integer, making use of an argument name, in PL/pgSQL:
249
250 CREATE OR REPLACE FUNCTION increment(i integer) RETURNS integer AS $$
251 BEGIN
252 RETURN i + 1;
253 END;
254 $$ LANGUAGE plpgsql;
255
256
257 Return a record containing multiple output parameters:
258
259 CREATE FUNCTION dup(in int, out f1 int, out f2 text)
260 AS $$ SELECT $1, CAST($1 AS text) || ' is text' $$
261 LANGUAGE SQL;
262
263 SELECT * FROM dup(42);
264
265 You can do the same thing more verbosely with an explicitly named com‐
266 posite type:
267
268 CREATE TYPE dup_result AS (f1 int, f2 text);
269
270 CREATE FUNCTION dup(int) RETURNS dup_result
271 AS $$ SELECT $1, CAST($1 AS text) || ' is text' $$
272 LANGUAGE SQL;
273
274 SELECT * FROM dup(42);
275
276
278 Because a SECURITY DEFINER function is executed with the privileges of
279 the user that created it, care is needed to ensure that the function
280 cannot be misused. For security, search_path should be set to exclude
281 any schemas writable by untrusted users. This prevents malicious users
282 from creating objects that mask objects used by the function. Particu‐
283 larly important in this regard is the temporary-table schema, which is
284 searched first by default, and is normally writable by anyone. A secure
285 arrangement can be had by forcing the temporary schema to be searched
286 last. To do this, write pg_temp as the last entry in search_path. This
287 function illustrates safe usage:
288
289 CREATE FUNCTION check_password(uname TEXT, pass TEXT)
290 RETURNS BOOLEAN AS $$
291 DECLARE passed BOOLEAN;
292 old_path TEXT;
293 BEGIN
294 -- Save old search_path; notice we must qualify current_setting
295 -- to ensure we invoke the right function
296 old_path := pg_catalog.current_setting('search_path');
297
298 -- Set a secure search_path: trusted schemas, then 'pg_temp'.
299 -- We set is_local = true so that the old value will be restored
300 -- in event of an error before we reach the function end.
301 PERFORM pg_catalog.set_config('search_path', 'admin, pg_temp', true);
302
303 -- Do whatever secure work we came for.
304 SELECT (pwd = $2) INTO passed
305 FROM pwds
306 WHERE username = $1;
307
308 -- Restore caller's search_path
309 PERFORM pg_catalog.set_config('search_path', old_path, true);
310
311 RETURN passed;
312 END;
313 $$ LANGUAGE plpgsql SECURITY DEFINER;
314
315
317 A CREATE FUNCTION command is defined in SQL:1999 and later. The Post‐
318 greSQL version is similar but not fully compatible. The attributes are
319 not portable, neither are the different available languages.
320
321 For compatibility with some other database systems, argmode can be
322 written either before or after argname. But only the first way is
323 standard-compliant.
324
326 ALTER FUNCTION [alter_function(7)], DROP FUNCTION [drop_function(l)],
327 GRANT [grant(l)], LOAD [load(l)], REVOKE [revoke(l)], createlang [cre‐
328 atelang(1)]
329
330
331
332SQL - Language Statements 2008-06-08 CREATE FUNCTION()